Introduction

Microvascular decompression (MVD) is popularized as a potentially curative treatment for hemifacial spasm (HFS) [1,2,3]. The offending arteries are commonly located on the root exit zone (REZ) of the facial nerve [4]. A secure transposition of the offender can relieve symptoms in most patients [5, 6]. The conventional retrosigmoid approach is commonly adopted to access the cerebellopontine angle, where the medial edge of the sigmoid sinus is exposed, and the dura is reflected along with the edge of the sinus to create an operative corridor [7, 8]. Irrespective of an appropriate retrosigmoid craniectomy, there are some difficult cases in manipulating neurovascular compression (NVC) due to a narrow operative field. In such cases, MVD may fail due to missing the offender or incomplete transposition of the responsible vessels [9]. Manipulating the arteries and cranial nerves in a narrow surgical field carries the risk of serious neurological complications. The surgical corridor is mainly affected by the site of the NVC, the size of the flocculus, and the location of the sigmoid sinus. Obtaining adequate exposure is essential to performing a safe surgery as well as avoiding risks.

Skull base approaches, such as the extended retrosigmoid approach [10,11,12] or the retrolabyrinthine presigmoid approach [13,14,15,16] are used to widen the operative field for medially located cerebellopontine angle lesions. With the extended retrosigmoid approach, the sigmoid sinus is thoroughly skeletonized with partial removal of the presigmoid air cells to retract the dura laterally together with the sigmoid sinus [10, 11]. The presigmoid approach allows observation far laterally with the retrolabyrinthine mastoidectomy to create a surgical corridor [13, 15, 16]. These approaches are rarely discussed in MVD surgery [17, 18]. We have been applying these approaches to patients whose surgical field is assumed to be narrow. The aim of this study is to find an indicator of when to apply a skull base approach by retrospective evaluation of our approach selection.

Methods

Patient cohort

There was a total of 28 patients (8.4%) in whom the extended retrosigmoid approach (n = 27) or the retrolabyrinthine presigmoid approach (n = 1) was applied among 335 consecutive MVDs for HFS from April 2005 to May 2022 (Fig. 1). These approaches were used for patients whose operative field was predicted to be narrow due to a medially located NVC, a large flocculus, or a medially located sigmoid sinus. Patient demographic data was collected from the medical records. There were 19 females (68%) and 9 males (32%), and the mean age at surgery was 54 years, ranging from 23 to 73 years. All patients suffered from typical HFS that affected one side of the face, except for one patient who suffered bilateral HFS due to tortuous vertebral arteries on both sides. The left side was more dominantly affected in 19 patients (68%) than the right side in 9 patients (32%). The duration between the onset and MVD was a median of 5 years and a mean of 5.4 years, ranging from 1 to 18 years. Five patients (18%) had previously failed MVD(s) (Table 1).

Fig. 1
figure 1

Participant flow diagram. MVD, microvascular decompression; HFS, hemifacial spasm. The skull base approaches include the extended retrosigmoid approach (n = 27) and the retrolabyrinthine presigmoid approach (n = 1)

Full size image
Table 1 Summary of patient characteristics, imaging analyses, and operative results
Full size table

Imaging study

All patients were examined preoperatively with computed tomography (CT) and magnetic resonance imaging (MRI) in fast imaging employing steady-state acquisition (FIESTA) (0.8-mm-thick) and contrast-enhanced T1 spoiled gradient recalled (SPGR) (0.6-mm- thick) images. Three-dimensional (3D) images of the operative view were created using the GammaPlan (Elekta, Stockholm, Sweden) by drawing the anatomical structures related to this procedure [19]. CT with bone window images was taken in all patients postoperative day one.

Measurement of the flocculus height and sigmoid-petrous distance

Exposure of the REZ is mainly limited by the site of the NVC, the size of the flocculus, and the location of the sigmoid sinus for the dural opening. The following two measurements were performed on FIESTA or T1 SPGR images for patients who had the skull base approach applied (n = 28) and for those who used the conventional retrosigmoid approach for a reference (n = 290). A line connecting the NVC to the medial edge of the sigmoid sinus is drawn to demonstrate the approach direction. The most distant point on the flocculus from the line is identified, and the perpendicular length to the line is measured as the flocculus height (FH, Fig. 2a). The distance between the posterior border of the sigmoid sinus and the petrous surface is measured at the level of the jugular foramen as the sigmoid-petrous distance (SP, Fig. 2b). The data set is plotted into a coordinate with FH value on the X-axis and SP value on the Y-axis. The standard deviation value of FH and SP among all patients is calculated for each patient. In describing the measured values, P-values were obtained by linear regression test. The statistical test was set to be significant at P < 0.05 (2-sided P-value). SAS software (version 9.4; SAS Institute, Inc., Cary, North Carolina) was used for the statistical analyses.

Fig. 2
figure 2

Measurement method. a The flocculus height (FH) is measured as the perpendicular length from the most distant point on the flocculus to the line (yellow) connecting the NVC (neurovascular compression, white arrowhead) to the medial edge of the sigmoid sinus (red arrowhead). b The sigmoid-petrous length (SP) is measured at the posterior edge of the sigmoid sinus from the line (white) starting from the midpoint on the clivus along with the petrous surface (white arrows) at the level of the jugular foramen (green arrowhead)

Full size image

Operative technique

All patients are operated in the lateral position with monitoring of auditory brainstem evoked response (ABR) and lateral spread response (LSR). When the amplitude of ABR decreases to less than 50% of baseline, the operative procedures are halted until ABR recovers to more than 70% of baseline for hearing preservation. LSR is used to confirm the adequacy of nerve decompression [20]. Spinal drainage is placed for patients whose subarachnoid space is assumed to be narrow. A skin incision is made at the retro auricular region, and the suboccipital muscles are divided between the occipitomastoid suture and the occipital artery. A craniectomy is performed using either the conventional or the extended retrosigmoid approach except for one patient with a retrolabyrinthine presigmoid approach. In the conventional retrosigmoid approach, the craniectomy is limited to exposing the medial edge of the sigmoid sinus (Fig. 3a). In the extended retrosigmoid approach, the craniectomy is extended to skeletonize the sigmoid sinus toward the occipital condyle. A limited posterior mastoidectomy is added to create room to allow for lateral mobilization of the sinus. (Fig. 3b). Partial far lateral transcondylar drilling to flatten the posterior fossa base allows access toward the REZ beyond the lower cranial nerves [21]. In the retrolabyrinthine presigmoid approach, the lateral aspect of the posterior fossa dura is widely exposed by the mastoidectomy (Fig. 3c). The dural incisions differ in each approach. While the upper end of the incision is placed near the mastoid emissary vein in the conventional retrosigmoid approach, it comes to the transverse-sigmoid junction in the extended retrosigmoid approach. In the presigmoid approach, the dural incision is made on the lateral aspect of the posterior fossa dura along with the anterior edge of the sigmoid sinus toward the jugular bulb (Fig. 3a’, 3b’, 3c’). The different approach direction to the REZ is shown in figure (Fig. 3a’’, 3b’’, 3c’’). After the dural incision, the cerebellar hemisphere is gently retracted from the base and the lateral surface of the posterior fossa. The lower cranial nerves are identified. After elevating the flocculus and the choroid plexus of the lateral recess, the offending vascular loop is usually found above the glossopharyngeal nerve. The REZ of the facial nerve is confirmed with a facial nerve stimulator. The offending vessels are transposed from the REZ with a Teflon sling or bridge as described in our previous report [22]. Fibrin glue is applied to fixate the Teflon sling onto the petrous dura. Adequacy of nerve decompression was confirmed by thorough exposure of the REZ and disappearance or decreased amplitude of LSR [20]. When a normal muscle response wave persists after decompression, no offending vessel on the REZ is re-confirmed using a facial stimulator. The dura is closed with the double collagen matrix grafting (DuraGen®, Integra Lifesciences, Plainsboro, NJ, USA) [23]. The air cells are sealed with small pieces of collagen matrix grafts with fibrin glue. Calcium phosphate paste (BIOPEX-R®, HOYA Technosurgical, Tokyo, Japan) with vancomycin is used for the cranioplasty.

Fig. 3
figure 3

Craniectomy and dural incisions. a, a’, a’’ The conventional retrosigmoid approach. The craniectomy is limited to exposing the medial edge of the sigmoid sinus (a). The dural flap is inverted along with the medial edge of the sigmoid sinus from near the mastoid emissary vein toward the lowest point of the sinus. The reflection of the dural flap does not affect the location of the sigmoid sinus (a’). A schematic axial slice shows the corridor to the NVC through the conventional retrosigmoid approach (a’’). b, b’, b’’ The extended retrosigmoid approach. The craniectomy is extended to skeletonize the sigmoid sinus (b). Inversion of the dural flap together with the sigmoid sinus provides a wider corridor and more medial access than the conventional retrosigmoid approach. The presigmoid mastoid air cells are drilled in cases where more lateral retraction is required. Note that the sigmoid sinus is almost collapsing by the dural retraction (b’). A schematic illustration shows a more lateral approach direction to the NVC (b’’). c, c’, c’’ The retrolabyrinthine presigmoid approach. An extensive mastoidectomy is performed to expose the lateral aspect of the posterior fossa dura. Bone removal reaches the temporal tegmen superiorly and the infrajugular condylar fossa inferiorly. The retrosigmoid area is partially opened to expose the entire sigmoid sinus (c). The dural incision is made on the dura along with the anterior edge of the sigmoid sinus (c’). A schematic axial slice shows a far lateral approach to the NVC (c’’)

Full size image

Assessment of surgical outcomes

For twenty-eight patients with application of the skull base approach, their postoperative neurological status was assessed either at our clinic, by telephone interviews, or by mailed questionnaires for remote patients. The status of facial spasms was evaluated immediately after surgery, postoperative 1-week, 1-month, 1-year, and the last follow-up visit. Any neurological complications, including delayed facial palsy, post-operative facial weakness, hearing impairment, hoarseness, dysphagia, cerebrospinal fluid leakage, and infectious complications were recorded.

Results

FH and SP measurements

The measurement results from 28 patients with skull base approaches are demonstrated (Table 1). The FH is measured at a median of 11.2 mm and a mean of 10.9 mm, ranging from 4.2 to 14.8 mm. The standard deviation value of FH in the patients with the skull base approach is a median of 62.0 and a mean of 61.1 among all patients. The SP is measured at a median of 6.6 mm and a mean of 6.5 mm, ranging from 0 to 13 mm. The standard deviation value of SP is a median of 52.1 and a mean of 51.6, ranging from 23.3 to 80.0. The standard deviation value is larger in the FH measurement (the median 62.0) than in the SP measurement (the median 52.1). The range is narrower in the FH measurement (40.7–73.1) than in the SP measurement (23.3–80.0).

The graph plotting the values of FH and SP from all patients is shown (Fig. 4). The measurement values from the patients with the skull base approaches are shown in red. The values from the patients with the conventional retrosigmoid approach are indicated in blue. The averages are shown as large dots, respectively. Most patients with the skull base approaches were distributed in the right half of the coordinate which indicates the FH value is larger than those of the conventional retrosigmoid approach. The SP measurement values vary on Y-axis in both approaches.

Fig. 4
figure 4

Distribution of the FH and SP values. The measurement result from all patients is plotted in a coordinate with FH value on the X-axis and SP value on the Y-axis. The small red dots represent the cases with the skull base approaches applied. The small blue dots indicate the cases with the conventional retrosigmoid approach used. The large dots are the average of both approaches. A horizontal line (green) and a vertical line (yellow) indicate the average value of the SP and FH in the conventional retrosigmoid approach. Most cases with the skull base approach (red dots) show a larger FH value than the average of the conventional retrosigmoid approach (yellow line)

Full size image

Analyses of the standard deviation and chart plotting show that the FH measurement is more consistent with our choice of skull base approach than the SP measurement. Statistical analyses showed the FH measurement has significant differences between both approaches (Fig. 5).

Fig. 5
figure 5

Comparison of the FH value. The FH measurement in the skull base approach demonstrated larger values than in the conventional retrosigmoid approach with a statistically significant difference (P < 0.0001)

Full size image

Operative results

The extended retrosigmoid approach was applied in all patients except one (96%). Intraoperative spinal drainage was placed in 3 patients (10.7%) for patients with a narrow subarachnoid space. The REZ of the facial nerve was safely exposed and the offenders were securely transposed in all patients. The offenders involved were either single or multiple vessels. The posterior inferior cerebellar artery (PICA) and the anterior inferior cerebellar artery (AICA) were similarly involved in NVC in 12 patients (43%), respectively, followed by the vertebral artery (VA) and the common trunk of the anterior and inferior cerebellar artery (APC) in 8 patients (29%) each. The transposition was performed using either a Teflon bridge in 13 patients (46%), a Teflon sling in 11 patients (39%), and a Teflon wedge in 4 patients (14%). A decrease in the ABR amplitude was noted in 2 patients (7.1%), which was transient and recovered by the end of surgery (Table 1).

Surgical outcomes and follow-up

The follow-up period was a median of 35 months and a mean of 40 months, ranging from 1 to 151 months. Immediate spasm-free was noted in 23 patients (82%). They decreased to 19 patients (68%) at postoperative 1-week. At the final follow-up, 27 patients (96%) were noted as spasm-free. One patient (3.6%) recurred 2 years after the first MVD, in whom exploratory surgery was conducted and an unrecognized offender was revealed at the mid-cisternal portion. Delayed facial palsy was observed in 2 patients (7.1%), which resolved within a month. Immediate facial weakness was found in one patient (3.6%) who required extensive manipulation during the decompression procedure. No persistent neurological deficit was found postoperatively. One patient had a CSF leak (3.6%). Meningitis was noted in one patient (3.6%). One patient had a wound infection (3.6%) in total (Table 1).

Case presentations

Case 1: The extended retrosigmoid approach for a repeat MVD (Fig. 6)

Fig. 6
figure 6

The extended retrosigmoid approach for the 3rd MVD (49 M, left). a A measurement of FH value. The FH is a perpendicular length measured from the most distant point on the flocculus to the line connecting the NVC and the medial edge of the sigmoid sinus. b An operative photo in the second MVD through the conventional retrosigmoid approach. A Teflon mass inserted in the previous MVD disturbed the access to the NVC. c The extended retrosigmoid approach in the third MVD allowed exposing the NVC. The posterior inferior cerebellar artery (PICA) is compressing the root exit zone (REZ) of the facial nerve. d A postoperative computed tomography (CT) shows extended drilling of the mastoid air cells over the sigmoid sinus

Full size image

A 49-year-old man suffered from HFS on the left for 5 years. The FH value in this patient was 14.2 (Fig. 6a). His first MVD was operated elsewhere through the conventional retrosigmoid approach with a padding method using Teflon felt. His HFS did not improve and he then was referred to our institution. An MRI revealed a PICA compression remaining on the REZ with deep indentation into the pons. The second MVD was conducted through the more inferior retrosigmoid approach. A Teflon mass was found around the mid-cisternal portion of the facial nerve root. The inserted Teflon felt was partially removed toward the REZ (Fig. 6b). Due to a narrow operative field, the PICA was transposed laterally at the descending portion without confirming the REZ, resulting in a failed second MVD. Postoperative MRI revealed the PICA loop remained on the REZ. Reviewing a 3D image from an operative view suggested that the decompression procedure was difficult through the conventional retrosigmoid approach due to a large flocculus and the medially located PICA loop. The third MVD was conducted 1 week after the second surgery with the extended retrosigmoid approach. The PICA loop compressing the REZ was clearly visualized (Fig. 6c). Decompression was performed safely. The patient became spasm-free immediately after surgery and maintained it for 3 years. No postoperative complication was noted. The postoperative CT shows sufficient drilling of the mastoid air cells over the sigmoid sinus to gain an operative view from the lateral aspect (Fig. 6d).

Case 2: The extended retrosigmoid approach for the first MVD (Fig. 7)

Fig. 7
figure 7

The extended retrosigmoid approach for the 1st MVD (54F, left). a The FH value was 12.4 mm, which predicts a narrow operative field if operated through the conventional retrosigmoid approach. b A three-dimensional (3D) image shows the relationships between the cranial nerves and the offending vessels. The vertebral artery (VA) and the PICA are compressing the pons deeply at the REZ of the facial nerve c An operative photo showing presigmoid bone drilling after skeletonizing the sigmoid sinus (SS, sigmoid sinus; TS, transverse sinus). The presigmoid dura is seen under the suction tube. d The dura is inverted together with the sigmoid sinus colored in purple. The venous flow is confirmed with Doppler ultrasound from the intradural side (red arrow). e An operative photo showing the exposure of the offending vessels and the lower cranial nerves. The NVC on the facial nerve is located on the other side of the offender. f A bone window CT slice shows the extended drilling of the mastoid air cells over the sigmoid sinus along with the lateral wall of the petrous bone

Full size image

A 54-year-old woman with HFS on the left was operated through the extended retrosigmoid approach for the first MVD based on the FH measurement. The FH value was 12.4, predicting a narrow surgical corridor if operated through the conventional retrosigmoid approach. (Fig. 7a). A 3D image depicted a deeply indented VA and PICA into the pons (Fig. 7b). A partial mastoidectomy with infrajugular drilling was performed after skeletonizing the sigmoid sinus to allow lateral retraction of the sinus (Fig. 7c). A dural incision from the transverse-sigmoid junction toward the inferior point of the sigmoid sinus was made. The dural flap was inverted together with the sigmoid sinus to create a wide operative view from the lateral aspect. The venous flow of the sigmoid sinus was confirmed patent with Doppler ultrasound (Fig. 7d). The offending arteries were clearly exposed (Fig. 7e) and safely transposed using a Teflon bridge. Her facial spasm disappeared immediately after surgery without any complications. No recurrence was noted for 14 months. A postoperative CT shows extensive drilling of the mastoid air cells of the presigmoid area (Fig. 7f).

Case 3: The retrolabyrinthine presigmoid approach for a repeat MVD (Fig. 8)

Fig. 8
figure 8

The retrolabyrinthine presigmoid approach for the 3rd MVD (62F, left). a A 3D image showing the anatomical relationships of the related structures. The REZ is compressed by the PICA loop far medially and rostrally in the deep. Two Teflon felts are inserted around the VA in the previous two failed MVDs. The lower cranial nerves are dislocated with the tortuous VA rostrally. The retrolabyrinthine presigmoid approach was applied. b The NVC was located far medially and rostrally into the pons. The measurement of FH value (14.8 mm) indicates the conventional retrosigmoid approach may cause difficulty in exposing the NVC. c An operative photo after mastoidectomy. The presigmoid dura, the posterior semicircular canal (PSC), and the endolymphatic sac (ES) are exposed. Infrajugular drilling was extended toward the condylar fossa. The retrosigmoid dura is partially exposed to allow medial retraction of the sigmoid sinus. The dural incision is made on the presigmoid dura (red dashed line). d An operative photo after dissecting the previously inserted Teflon felts from the lower cranial nerves. The deeply situated PICA loop is observed compressing the REZ of the facial nerve. e A postoperative CT shows the mastoidectomy reaches the labyrinth

Full size image

This is the only patient in whom the presigmoid approach was utilized (Fig. 8a). The FH value was 14.8, which was the largest in our series (Fig. 8b). A 62-year-old woman had suffered from severe HFS on the left for 7 years. Her previous two MVDs operated elsewhere resulted in failures due to incomplete decompression through the conventional retrosigmoid approach. Multiple Teflon felts around the lower cranial nerves and the facial vestibulocochlear nerve complex were found on radiological study. The location of NVC by a PICA loop was located far medially on the pons (Fig. 8a). An inferior retrosigmoid craniectomy and limited extreme lateral infrajugular transcondylar exposure (ELITE) approach was performed and the presigmoid dura was exposed toward the jugular bulb (Fig. 8c). Meticulous dissection and removal of Teflon felt from the cranial nerves allowed access to the REZ of the facial nerve through the presigmoid approach (Fig. 8d). The compressing PICA loop was visualized and transposed safely with Teflon slings. Her facial spasm was relieved immediately after surgery without any neurological complications. She maintained spasm-free for 5 years. A postoperative CT shows an extensive mastoidectomy for the retrolabyrinthine presigmoid approach (Fig. 8e).

Discussion

The retrosigmoid approach is generally applied in MVD for HFS [1, 5, 7, 8]. This approach provides a very sufficient operative field to manage almost all patients with HFS. It is rare that the REZ is inaccessible via the retrosigmoid approach except when the NVC is located far medial or rostral due to severe compression by the tortuous arteries [4, 24]. Some variations of the retrosigmoid approach in MVD surgery are reported to minimize cerebellar retraction and obtain sufficient exposure [25,26,27]. Extending the craniectomy inferiorly or laterally contributes to widening the operative corridor, however, the necessity for each patient has not yet been discussed. The significance of this study is to identify the necessity of additional drilling to the lateral before surgery. A larger FH value may predict when to consider the extended retrosigmoid approach.

Despite a variety of approaches in development to access medially located tumors and vascular lesions in skull base, the application of these approaches is rarely discussed in MVD surgery. The location of the REZ is usually accessible via the conventional retrosigmoid approach in most cases [1, 5, 7, 8]. Even in cases with far medially located NVC, the decompression procedures may be quickly accomplished with more retraction of the cerebellum for a limited time. Surgical maneuvers in a narrow operative field, however, involve risks such as cranial nerve or vascular injuries, and failure of MVD [9]. Obtaining a maneuverable operative field is essential to performing a safe and secure MVD. The skull base approaches should be taken into consideration when such operative conditions are predicted. Some authors reported the benefit of either fully endoscopic MVD [28, 29] or simultaneous endoscopic and microscopic visualization in MVD [30]. Endoscopic MVD minimizes the risks of brain retraction and extensive dissection required for microscopic procedures. Using an endoscope has an advantage over a microscope in the visualization of structure and identification of neurovascular conflict. Our simple method to predict a narrow operative field is also useful for considering the application of endoscopic MVD.

A variety of decompression methods for the facial nerve were reported in the literature [31,32,33,34,35,36]. The padding method using a shredded Teflon felt on the REZ to separate the offender was initially introduced [1,2,3]. Later, transposing the offender away from the REZ using a Teflon sling is reported to provide better surgical outcomes than the padding method [37]. While the padding method can be accomplished in a narrow operative field, the transposition technique with a sling requires a wider surgical field to manipulate the involving vessels. Visualizing the REZ and identifying the adjacent structures is the prerequisite for performing this technique. Limited exposure of the sigmoid sinus leaves a ledge of bone which compromises the surgical exposure and requires more cerebellar retraction to expose the REZ [11]. This study demonstrated that some patients could benefit from the skull base approach by widening the operative field, especially for repeat MVD cases.

The indication for a skull base approach should be carefully considered because unnecessary drilling of the mastoid air cells carries surgical risks, such as venous sinus injury, CSF leak, and wound infection [38]. One of the unique risks of the extended retrosigmoid approach is venous sinus occlusion. Tense retraction of the dura may cause sinus occlusion during surgery that may result in fatal sequelae due to venous infarction if it is the dominant side. Intraoperative confirmation of the venous flow using Doppler ultrasound or indocyanine green venography is useful for avoiding sinus occlusion. On the other hand, endoscopic MVD may avoid a mastoidectomy that can put the sigmoid sinus at risk or require more reconstructive steps if it is damaged [28,29,30]. It may also benefit in avoiding CSF leaks. Applying a skull base approach or an endoscopic procedure should be carefully considered based on the surgeon’s experiences and techniques.

This study demonstrated the anatomical diversity of the structures that relate to performing an MVD. Although skilled neurosurgeons may be able to manage HFS patients solely through the retrosigmoid approach, predicting the difficulty of the procedure and considering the additional bone removal to avoid a narrow surgical field contribute to performing a safe MVD, which may be more beneficial for less experienced neurosurgeons. Training on cadaver dissection and being familiar with petrous bone anatomy is necessary to take these approaches into surgical consideration.

Limitations of the present study include those that are inherent to studies of retrospective design with a small number of patients and a relatively short follow-up period. Another limitation is the comparison of two approaches on the same patient is impossible, therefore, the exact benefit of the skull base approach cannot be precisely evaluated. The absolute FH value cannot be generalized because there is an inherent bias in the FH measurement.

Conclusion

A simple preoperative measurement is introduced to evaluate the necessity of expanding a craniectomy laterally to obtain a wider operative field in MVD.